The Effects of Cognitive Style and Knowledge Structure on
Performance Using a Hypermedia Learning System
DR. ROCCO PAOLUCCI
Department of Computer Information Science
Cabrini College, Radnor, PA 19087
ABSTRACT
Hypermedia software is quickly becoming popular with many educational institutions and
curriculum programs. One of its major attractions is that it allows for the flexible
structuring, construction, and exploration of the knowledge domain. The purpose of this
research study was to determine the relationship among the cognitive style (active and
reflective), the structuring of the knowledge domain as controlled by the hypermedia
software (hierarchical, branching, and conventional), and the test performance of the
learner (in terms of cognitive skill levels). More specifically, a controlled group
experiment with 115 pre-adolescent, elementary school students (fifth-graders) was
conducted. An immediate written posttest, used to assess the subjects' learning
performance in terms of total, higher-order, and lower-order cognitive skills, was
administered. Significant differences were found among the three knowledge structure
schema groups for the total and the higher-order cognitive skills performance scores. No
significant relationship was observed between cognitive style and performance, nor did
this variable significantly interact with the knowledge structure variable. The major
conclusion from this study is that a relationship exists between the structuring of the
knowledge domain, as reflected by the hypermedia software, and positive learning
performance. This relationship is also extended to the effective use of higher-order
cognitive skills.
INTRODUCTION
Information technology has historically played a central role in the delivery of
instruction in the classroom. Over the years, teachers have used books, television,
projectors, and many types of lab equipment as tools to help them transfer knowledge to
their students (Hawkins, 1993). In the past decade, computers have been added to the
teacher's technology toolbox. Like no other previous technology, it has been able to
capture the imagination of educators (Soloway, 1993). Today, computers can be
frequently found in classrooms and laboratories throughout our schools, universities, and
other educational and training institutions.
Experience and previous research on the effectiveness of computer technologies
have consistently revealed that, when used appropriately, computers make excellent
learning tools (Snider, 1992). In the classroom, the applications of computers have
evolved from the provision of drill and practice for remediation, to later providing
structured curriculum and instruction. Today, the computer is often used for knowledge
exploration and construction. Often quoted benefits include, among others: Learner
pacing of instruction, interactive quality, immediacy of feedback, and learning
individualization (Hawkins, 1993).
Of these attributes, individualization and interactivity have commonly been
viewed as central in making the computer the learning tool of choice in the classroom of
the future. Specifically, it seems that by carefully choosing the appropriate software, the
2
computer has the inherent flexibility to accommodate a variety of student learning styles
(Hawkins, 1993; Schank, 1993). Furthermore, its interactive nature seems to be ideal for
exploring knowledge (Hawkins, 1993; Oren, 1990). Computing technologies have had
such a significant impact in education that many people believe the computer has the
potential to transform today's teacher-centered classroom into tomorrow's learnercentered classroom. This transformative role of technology is seen to be at the top of the
agenda of many educational reform initiatives (Hunter, 1993; Pearlman, 1993; Sheingold,
1991; Snider, 1992).
On the technological front, during the past decade, the computer has evolved from
being a command-line instructional machine. Graphical and friendlier user interfaces
have made human-computer interaction much easier and more effective. Today,
computers inexpensively and easily provide highly interactive multimedia information
(text, sound, images, and video) to its users. Furthermore, these interactive multimedia
systems may now be programmed to deliver hypermedia (Dede,1992).
Hypermedia software stores information in networks of nodes connected
by links (Ambrose, 1991). The nodes can contain text, graphics, audio, data, video, or any
other form of media. To the user, hypermedia programs are nothing more than knowledge
databases guided by menus and/or icons (Locatis et al, 1992). The essence of hypermedia
is the instantaneous and dynamic linking of concepts within the various knowledge
3
domains, resulting in a series of nonlinear connections. To facilitate navigation, an
interface or "browser" is provided for moving around the knowledge base.
Hypermedia is the result of the evolution of its predecessor, hypertext (nodes
contain only textual information). Over the years, this technology has evolved to
incorporate many types of media. Today, hypermedia is commonly available with many
different types of computer systems (PC, Mac, and others) and formats (floppy disk, CDROM, CD-I, etc.). However, the most recent hypermedia development effort could
potentially be the most revolutionary. The integration of hypermedia technologies with
high-speed computer networks has made the development of the Internet's World Wide
Web possible (Soloway, 1995). The wide-reach of the traditional Internet network (global
in scope), coupled with the multimedia capabilities of the computers it interconnects,
have made the World Wide Web a truly international and highly distributed hypermedia
system. This new and exciting form of hypermediated network is widely used in
education, and is becoming a popular tool for creating a self-directed and exploratory
learning environment within many classrooms (Dern, 1993, Soloway, 1995).
4
REVIEW OF THE LITERATURE
Over the past decade, a number of studies have explored the effects of hypermedia
on learning. A review of the literature by Ambrose (1991) has revealed a few interesting
research questions. The more pertinent and interesting ones for this research study are:
What type of learner is most likely to effectively utilize a hypermedia environment? Do
students using hypermedia materials learn more effectively than those using
more structured traditional material (i.e. using a more linear medium)? Does the use of
hypermedia materials encourage learners to employ higher-level thinking skills?
Furthermore, a research review study of hypertext by Tsai (1989) also raised some
similarly interesting questions. The more salient and pertinent of these is: Will learners
form better conceptual models usinghypermedia or will the richness of the information
overload learners and impair effectiveness and efficiency of learning?
In assessing the effects of hypermedia on learner performance, the above
questions can be classified as addressing the understanding of two major issues, learner
characteristics and learner control. This research study will examine the empirical studies
associated with these issues. The research will address the effects of the learner's
characteristics (by specifically examining cognitive style) and the learner's level of
5
self-direction (by addressing the structuring impact of the knowledge domain) on the
performance of students when operating in a hypermedia learning environment. The
performance will be examined by evaluating the use of higher-level cognitive skills
as well as the overall achievement test scores. The research which address the above two
factors (cognitive style and knowledge structure) will be more fully described in the
following sections.
Cognitive Style
Although used interchangeably in the literature, there is a technical difference
between the use of the terms cognitive style and learning style. Cognitive style deal with
the "form" of cognitive activity (i.e., thinking, perceiving, problem solving, etc.), not its
content . It is viewed to be "persuasive dimensions" of personality, bipolar in nature, and
stable over time (Whyte, Karolick, Nielsen, Elder, & Hawley, 1995). Learning style, on
the other hand, is seen as a broader construct, which includes cognitive along with
affective and physiological styles (Keefe, 1988). Therefore, the distinction being made, it
should be clearly stated that this research study will deal with cognitive style, even though
at times the term learning style will be used (mostly in reference to previous research).
6
When it comes to the use of hypermedia, what type of learner is most likely to
effectively utilize a hypermedia environment? A theoretical study review by Locatis et al
(1992) claims that:
Persons already sophisticated in the subject matter or who have high general
ability may learn most from hypermedia. Since sophisticates are likely to have
representations similar tothose authoring the knowledge, they may navigate links
and process information more appropriately....Lower ability students might benefit
if there is extra guidance. (p. 73)
Thus, Locatis et al (1992) seem to imply that learners having general ability (e.g.,
cognitive style) should perform well in a hypermediated environment. Overall,
hypermedia offers the potential for new strategies of learning, teaching, studying, and
creating. An experiment with 63 undergraduate students using hypermedia studied the
major effects of learning style preferences, GPA, and attitudes on learner achievement. It
showed that GPA had a significant and positive correlation with achievement. However,
this conclusion was observed only when GPA was included with the attitude scale into
the regression analysis (Billings & Cobb, 1992). According to the authors, their results
seem to contradict previous studies which found found no relationship at all between
attitude and achievement and an insignificant interaction between learning style, attitudes,
and achievement (Billings & Cobb, 1992).
7
Another field study of adult computer technicians located throughout the country
examined the relationship between learning style and the effectiveness of an interactive
video instructional system (Larsen, 1992). The results failed to detect any significant
correlation between learning style and achievement scores.
An experimental study by Yung-Bin (1992) tested the effects of learning style and
on-line instructional advisement onundergraduate students' achievement test performance.
The results indicate that achievement test scores were affected by the interaction of
learning style and advisement, with active-learning subjects tending to score higher on the
achievement test than passive-learning subjects. Furthermore, passive-learning subjects
who received the advisory version of the software scored significantly higher on their
achievement test than passive-learning subjects who had the non-advisory version.
Finally an experiment by Jonassen and Wang (1993) involving ninety-eight preservice teachers in an open enrollment college showed that individual differences
interacted with learning through the use of hypermedia. The research showed that the
"field independent processors" cognitive style subjects generally preferred to impose their
own structure on information rather than accommodate the structure that is implicit in the
materials. The authors conclude by stating that it is likely that field independent learners
are better hypermedia processors, especially as the form of the hypermedia becomes more
referential and less overtly structured (Jonassen & Wang, 1993).
8
In summary, the research, albeit mixed, indicates that learners with certain
cognitive styles (labeled active, exploratory, or field independent) should perform better
in a hypermedia learning environment than those with others. However, this assertion
seems to be more the case when cognitive style interacts with other factors, such as
attitude and advisement (Paolucci & Wright, 1996).
Knowledge Structure
Will learners form better conceptual models using hypermedia or will the richness
of the information overload learners and deteriorate effectiveness and efficiency of
learning? According to Marchionini (1988), "Although self-directed and exploratory
learning are worthy objectives to achieve in learning environments, freedom to learn does
not seem to be a sufficient condition to assure effective learning." (p. 11). The point to
be made here is that freedom can be confusing, since it increases decision-making load.
Attention may be diverted from content and relationships as learners attend to
navigational decision-making and disorientation may be experienced. This problem may
be compounded by the vast quantities of easily accessible information, much of which
may only be peripherally relevant.
9
Thus, it seems, the rich learning environment can easily become a trip into
"hyperspace" (Marchionini, 1988). However, Tsai (1989) claims that this issue can be
adequately addressed, "Even though users can freely choose their browsing paths, a
hyperdocument has an intrinsic structure which is determined by how the nodes are
linked together. This intrinsic structure should have some effects on users' browsing and
commenting activities." (p. 11). This is similarly noted by Conklin (1987), with his claim
that there is no natural topology for an information space.
These observations raise yet another interesting question: Are some knowledge
structure schemas better than others in facilitating and promoting learning in general, and
higher-cognitive skills in particular? Are there schemas that are better suited for certain
cognitive styles? Empirical studies on this subject have yielded different results. The
first investigation by Stanton (1994) examined the comparison of a learning environment
that presented the instruction knowledge in a linear format over which the student had no
influence, with another in which the student was able to determine the sequence
of the instruction freely. The study results, involving forty adult students, suggest that
performance may be improved in the non-linear learning environment, but this is likely to
be dependent upon the ability of the participants to organize the environment in a manner
that suits their own learning style. Stanton's findings also support the theoretical research
of Brooks, Simutis, and O'Neil (1985), who claim that learners may actually be able to
10
process information more effectively if it is presented in a manner that is closely matched
by their learning style.
However, a study by Jonassen and Wang (1993) involving ninety-eight preservice teachers in an open enrollment college indicated that merely showing learners
structural relationships is probably not sufficient to result in greater structural knowledge
acquisition or performance. The authors observe that:
It appears that arbitrarily imposed semantic nets are not adequate to overcome the
personal ones or at least not directly mapped onto the learners. What seems to
matter most is the construction of personally relevant knowledgestructures. (p. 7)
The authors claim that such constructions can promote higher-order thinking in the form
of logical reasoning. Furthermore, a study of health care employees by Cordell (1991),
designed to determine whether knowledge structure and learning styles affect outcome,
revealed insignificant and unclear results. Similarly, a study by Rasmussen and Davidson
(1996) with pre-service teachers as subjects did not yield a significant relationship
between hypermedia structure and performance; however, the results showed a significant
interaction between learning styles (active-reflective) and hypermedia structure
(hierarchical). Thus, although only a few empirical research studies exist on this topic,
there seems to be no clear empirical evidence substantiating the theoretical claim that the
structuring of the knowledge domain onto the hypermedia courseware, and its use,
has any relationship to learning performance.
11
Having examined the dimensions of cognitive style and knowledgestructure, the
questions of effectiveness in terms of performance skills remain. Does the use of
hypermedia materials encourage learners to employ higher-level thinking skills? Is there a
relationship between the hypermedia knowledge base structure and the promotion of
higher-order cognitive skills?
It has been a commonly held view that discovery learning environments are ideal
for stimulating higher-order thinking (Schank, 1993). These views have been discussed
and documented at length by such noted cognitive theorists as Bloom (1956) and Gagne
(1977). In using hypermedia environments to teach high-level learning skills, Locatis et al
(1992) claim that:
Hypermedia might best be used when discovery learning is warranted. Since
discovery implies teaching the process of learning as well as acquisition of
content, students might be encouraged to construct their own knowledge
representations and to refine and reflect upon them periodically in the course of
exploring a hypermedia knowledge base to solve problems -- a strategy analogous
to one used by some intelligent tutoring systems to ensure knowledge is properly
understood, integrated and remembered.(p. 73)
Much theoretical research in this area seems to imply that hypermedia improves learning
by focusing attention on the relationships of ideas, rather than isolated facts, facilitating
12
concept formation, understanding, and association (Ambrose, 1991). According to some,
in tracking the subject's navigational patterns when using a hyperdocument, the learner
constructs a knowledge map by specifying the relationships between concepts in the
domain (Nelson, 1992). The resulting map can be used to recognize the learner's
conceptual associations, in addition to recognize the learner's misconceptions (Kumar,
1994).
However, the empirical research on the assessment of higher-order thinking skills
using hypermedia software seems to indicate, but does not directly establish, such a
relationship (Jonassen and Wang, 1993). Moreover, the number of studies that buttress
this theoretical claims are almost non-existant.
METHODS
As hypermedia technologies become more commonly used in education, students,
teachers, instructional designers, and administrators need to understand how these can be
applied to enhance the instructional and learning processes. Toward this end, it is
important for those teachers who wish to be local producers of hypermedia courseware
(through the use of authoring software) to understand the implications of structuring
knowledge in their design of the hyperlessons (Cates, 1992; Tolhurst, 1992).
13
It is equally important to identify those learner characteristics or factors (e.g.,
cognitive styles) that foster effective use of the hypermedia systems, so that teachers can
match and accommodate these to the proper learning environment. Finally, it is
imperative for teachers and administrators to understand how the structuring of
hypermedia, if done well, can foster and augment a variety of student learning skills.
The purpose of this research study is to evaluate the effects of cognitive style and
the structuring of the knowledge domain, as represented by the hypermedia courseware,
on the learning performance of young adolescent elementary school students.
Research Design
An experiment was performed with six classrooms of early adolescent elementary school
students (fifth-graders). The experiment took place near the end of the school year.
Hyperdocuments developed specifically for this study, using hypermedia authoring
software, was used as learning systems in delivering a short unit of instruction.
The subject matter used in the development of the hyperdocument lesson was in
the science domain, more specifically a lesson unit on "ecology". The choice was
influenced by both its relevance to the curriculum, and the high interest level that
14
has historically been shown by the students on this topic. However, and more importantly,
it should be noted that this topic was mainly chosen to control for the subjects' prior
knowledge, since this factor can have an impact on the effectiveness of the hyperlesson.
The objective was to choose a topic with which the subjects where somewhat familiar, in
order to provide a decent knowledge base to facilitate further learning. Moreover, the
knowledge base had to be somewhat equivalent for all participating subjects, in order not
to bias the performance outcomes of the experiment.
A lesson on Ecology met both requirements, all fifth-grade students in the
population had received a "hands-on" activity lesson on ecosystems in the beginning
of the school year, months prior to the research experiment. Given this, it is safe to
assume that the subjects' level of prior knowledge on the topic of ecology was relatively
the same, and hence not considered to be a factor on the outcome of the research study.
Furthermore, since all participating students had a familiarity with the basic concepts of
an ecosystem, the hypermedia lesson on ecology could build on this knowledge base
and facilitate additional learning.
The hypermedia software was programmed and structured to produce three
different hyperdocument options, each designed to reflect three different and widely used
schemas in the structuring of the knowledge domain. These are, conventional (the nodes
and links of the hyperdocument allow for significant navigational freedom and self-
15
direction), branching (the nodes and links of the hyperdocument allow for a moderate
degree of navigational freedom and self-direction), and hierarchy (the nodes and links
of the hyperdocument allow for little navigational freedom and self-direction). It should
be noted here that, although structured differently, all three hyperdocuments contained the
exact same information or knowledge base.
The six classrooms, and their respective student subjects, were randomly
subdivided into three different groups (each allowing the use of the three hyperdocument
versions). All six participating teachers were instructed to review the previously given
activity lesson on ecosystems, and relate this information to the use of the ecology
software. Students were also informed that they would have to take a written test upon
completion of the hyperlessons. To provide an incentive to perform well, the teachers
informed the students that the results of the test would be used for extra credit. After
receiving some preliminary instructions by the teachers and the researcher, the students
used the hyperdocument in a computer classroom for 40-45 minutes. All classrooms were
equipped with multimedia-ready Macintosh computers that easily allowed for the use of
the hyperdocuments.
The exposure and use of the hyperlesson was repeated approximately a week later;
once again, in the schools' computer classroom. A written performance posttest, designed
to evaluate the levels of student performance in terms of higher and lower level cognitive
16
skills, was administered by the participating classroom teachers and given to all students
immediately after the second exposure.
During this period, a written test was also administered to determine the students'
cognitive styles. It is worthwhile to note here that, in order to minimize the influence of
the research study on the normal classroom organization and procedures (and perhaps
outcomes), both tests were administered by the classroom teachers.
Finally, in order to get a baseline measurement of the effectiveness of the
hypermedia learning system, an additional fifth-grade classroom was chosen to participate
in the research study as a control group. Naturally, the student subjects in this control
group did not get to use the hyperlesson on ecology, but were asked to take the
written performance test.
Population and Sample
The sample for this study consisted of six, fifth-grade classrooms of students in a
suburban, middle-income school district. The classrooms and students were located in
three different elementary schools. All of the students and their respective teachers were
computer literate, deliberately chosen to control for this factor in the study. The original
total number of participating student subjects was 131.
17
The six classrooms of children were randomly assigned one of three specific
versions of the hyperdocument. In all, they formed three distinct groups, each reflecting
three schemas or levels of knowledge structure representations. This method was
chosen over other options, including the one where students in each classroom are
randomly subdivided into three groups, with each group using one of the three versions of
hyperdocuments. It was felt that having students using different versions within
the same classroom would promote discussion and comparison, potentially biasing,
contaminating, and invalidating the results (Borg & Gall, 1989).
Thus, each of the three hyperdocument versions was used by two teachers, each
expected to have an approximately equal number of student subjects. It was also expected
that each of the three hyperdocument versions would have an approximate equal number
of subjects representing the different cognitive styles.
Instrumentation
In performing the research experiment described above, four different instruments
were used to develop the necessary factors that formed the set of independent, covariate,
and dependent variables. These instruments are: The hyperdocuments, the cognitive style
18
assessment test, the Comprehensive Test of Basic Skills (CTBS), and the learning
performance test.
Hyperdocuments. The hyperdocuments used in this research study were developed
using HyperStudio(TM) multimedia authoring software for the Macintosh computer
platform (Wagner, 1995). Much research has been performed on techniques for authoring
and structuring hypermedia software (Kearsley, 1988). However, as shown in Figure 1,
three major design approaches or schemas have emerged from this work: Conventional,
Branching, and Hierarchy (Locatis et al, 1992; Misanchuk & Schwier, 1991).
With the conventional approach (see Figure 1a), most topics are linked
referentially on the assumption that students can begin learning some subjects when
requisite knowledge is only partially grasped. Students can freely access any of its nodes.
The nodes and links of the conventional hypermedia document are relatively
unstructured, allowing for a high degree of self-direction.
With the branching approach (see Figure 1b), the learner is presented with choice
points or nodes at which different responses will result in different alternative paths
through the instruction. In some cases, access to lower level nodes might be blocked until
all nodes at higher levels are viewed. The nodes and links of the branching
hyperdocument are somewhat structured, allowing for a medium degree of self-direction.
19
Finally, with the hierarchy approach (see Figure 1c), links are only established
between nodes containing information prerequisite to others. Users do not access nodes
teaching higher level skills until those at lower levels are accessed. The nodes and links
of the hierarchical hyperdocument are more tightly structured, allowing for a lower
degree of
self-direction (Locatis et al, 1992; Misanchuk & Schweir, 1991).
[Insert Figure 1 here]
At one extreme, the conventional hyperdocument will allow the student-user a
greater degree of freedom in exploring the knowledge domain. At the other extreme,
the hierarchical hyperdocument will allow the student-user less freedom in exploring the
subject's domain.
Thus, the hypermedia authoring software used in this research was programmed to
produce three distinct hyperdocuments (coined HyperEcology for the purpose of this
research study). These were developed to reflect the three distinct schemas or approaches
to knowledge structuring and representation described above. It is important to note that,
although the hyperdocuments were structured differently, their informational content or
knowledge database were identical.
Although it is very difficult to capture the essence of the three versions of the
ecology hyperdocuments without the use of a computer, it is valuable to illustrate their
schemas on paper. Figures 2 and 3 represent two of the three HyperEcology software
20
knowledge structures schemas, hierarchical and branching, respectively. Note, that the
conventional schema is not shown. The reason is due to the fact that the number of links
associated with this schema is fairly high. Thus, its figure would be confusing to visualize
and difficult to depict. To illustrate this point, the hierarchical hyperdocument is
comprised of 39 nodes and 37 links, the branching of 39 nodes and 50 links, while the
conventional of 39 nodes and 173 links.
[Insert Figures 2-3 here]
It also worthwhile to note that, for the hierarchical hyperdocument, each of the
screens was designed to allow the users to go back to the previous branching point (if one
existed). When the very bottom of the branch is reached, the user has the ability to go
back to the very top of the branch. The navigational design methodology for the
branching hyperdocument was exactly the same as the hierarchical, with the exception
that the user is allowed to branch out. However, in this case, the user cannot immediately
go back to the branching out point, but is instead provided with the choices of the new
branch. The links for the conventional hyperdocument were determined by the frequency
of the keyword titles associated with each node (e.g., the term "water" was linked to the
node "water" each time it appeared in any of the other nodes or screens). In essence, once
the program was started, the users of this schema did not have a way to go back to the
beginning of the program. Since with the conventional schema, there is no beginning and
21
end or top and bottom (although all the screens are connected and have many links among
them).
Therefore, it is important to note that while the hierarchical hyperdocument allows
the user to easily visit all the screens (nodes) with some opportunity for exploration but
no disorientation; the branching hyperdocument allows for more exploration and possible
disorientation (although, the user has the ability to overcome these events and visit all the
screens); the conventional hyperdocument allows for a significant degree of freedom for
self-exploration but also for significant disorientation. It is important to note that, with the
conventional hyperdocument, the fact that not all screens will be necessarily searched and
viewed by its user is a distinct possibility.
Cognitive Style Assessment. The instrument used in this research study to assess the
subject's cognitive style was the Kolb's Learning Style Inventory (Kolb, 1984). This
instrument has consistently been shown to possess both adequate reliability
and validity (Kolb, 1976; Smith & Kolb, 1986).
Kolb defines four learning styles, corresponding to each possible dimension
combination: Diverger (ability for viewing concrete situations from many different points
of view), Assimilator (ability to understand a wide range of information and putting it
into concise, logical form), Converger (ability for finding practical uses for ideas and
theories), and Accommodator (ability to execute plans and take risks). The theory claims
22
that students with different learning styles respond differently to the various instructional
strategies, and that these should be matched in achieving effective learning (Smith &
Kolb, 1986).
Kolb's learning style theory also claims that these four styles can differ and be
viewed along two different cognitive continuums or dimensions: Concrete ExperienceAbstract Conceptualization and Active Experimentation-Reflective Observation (Kolb,
1984). The first dimension is an assessment of the way people perceive information, with
some people preferring to sense and feel their way (Concrete Experience) while others
preferring to think their way through (Abstract Conceptualization). Along the second
dimension, how information is processed is assessed, with some people preferring to
jump in and try things (Active Experimentation) while others preferring to process new
information by reflecting on it (Reflective Observation).
According to the theory, the extremes of each dimension are mutually exclusive.
For example, it is claimed that individuals cannot simultaneously perceive by Active
Experimentation and by Reflective Observation, since these will produce a conflict
situation when presented with new information. The individual, in resolving this, must
choose how to perceive and process new information. Thus, each person develops a
preference by making frequent and successive choices in patterned ways.
The result of the above is a cognitive style, which is defined as the characteristic
way an individual generally chooses to perceive and process new information (Kolb,
23
1984). The Concrete Experience mode is included in the Diverger and Accommodator
learning styles, the Abstract Conceptualization in the Converger and Assimilator learning
styles, the Active Experimentation in the Converger and Accommodator learning styles,
and the Reflective Observation in the Diverger and Assimilator learning styles. It
is important to note that this research study is designed to focus on the information
processing (Active Experimentation and Reflective Observation), rather than the
perceptive dimensions of the Kolb's four learning styles.
Reading. Since the reading level is an important contributing factor towards the learning
performance outcomes of the student subjects, an instrument to assess this level was
required for this research study. The Comprehensive Tests of Basic Skills (CTBS) was
employed to accomplish this (CTB/MacMillan,1989). Reading is one of three sub-tests
(along with language and mathematics) that form this standardized achievement test.
Furthermore, the total reading sub-score has two components (vocabulary and
comprehension). The data used in this research study was the national total reading
percentiles scores of the participating student subjects. A percentile score indicates that
the student is performing at that level higher than other students who have taken the
CTBS nationally standardized test.
Performance Assessment. In order to evaluate the student subjects' performance after
24
using the hypermedia document, a written posttest was developed. This test was identical
for all student subjects, regardless of which hyperdocument version was used in the
experiment. The test is comprised of two cognitive sub-scores: Higher level skills
and lower level skills. The higher level cognitive sub-score questions consist of four
parts and is designed to assess the effects of the hyperdocument on the learners' structural
knowledge, while the lower level cognitive sub-score questions is designed to assess the
learners' informational recall skills.
One way of assessing higher-order thinking skills, as defined by Bloom (1956), is
by assessing structural knowledge (Dierkhoff, 1983; Jonassen, Beissner, & Yacci, 1993).
According to these researchers, structural knowledge is the knowledge of how
concepts within a domain are interrelated. It mediates the translation of declarative
knowledge into procedural knowledge. Questions used in testing this type of knowledge
are of four types: Analogies, associations, relationship proximity judgments, and semantic
relationships.
The analogies questions require students to assess similarities of concepts derived
from the knowledge domain. The association questions require students to choose the
strongest connection between two terms. The relationship proximity judgments questions
require that students choose one or more correct alternatives associated with a particular
idea, concept, or scenario. Lastly, the semantic relationships question requires students to
25
identify the nature of the relationship among terms or concepts by connecting these into
meaningful sentences. This part of the posttest consists of sixteen questions, for a
maximum total of twenty points (these equally distributed among the four types).
In addition to the questions used to assess higher-order structural knowledge,
fifteen multiple-choice information and recall related questions were developed to assess
lower-order cognitive skills. Each of these is worth one point, giving a fifteen for a
maximum score for this sub-score.
Thus, the total number of questions included in the posttest assessment instrument
is thirty-one for a maximum total score of thirty-five points, twenty of these from higherorder cognitive skills questions and fifteen from lower-order cognitive skills questions. It
is important to note that, what is of paramount significance to this research study is the
"relative" performance scores of the student subjects, rather than their "absolute"
performance scores.
Reliability and Validity
Hyperdocuments. All three hyperdocuments used in this research study were prototyped,
previewed, and evaluated in concert with a participating expert teacher to ensure
relevance to the curriculum and the appropriateness of the reading level of all textual
26
information. Furthermore, the prototypes were also evaluated by an academic expert in
science education. Modifications were made throughout the design process to reflect
both the experts' feedback on the subject domain and also to ensure the seamless
integration of the courseware into the curriculum's lesson plans.
Cognitive Style Instrument. Historically, several types of learning or cognitive style
instruments have been used to study the relationship between it and the effectiveness of
hypertext and hypermedia (Dunn, Dunn, & Price, 1984; McCarthy, 1981). However, the
Kolb Learning Style Inventory, was employed by this research study because it has
consistently been shown to possess both adequate reliability and validity (Kolb, 1976;
Smith & Kolb, 1986). Validity has been shown through significantly high correlations
with similar instruments, including the Myers-Briggs Indicator (Kolb, 1976).
Reading Level Instrument. The Comprehensive Test of Basic Skills (CTBS) instrument
is a historically established and respected standardized achievement test. It is widely
accepted by the basic educational community as being one of the most valid and reliable
indicators of curriculum and student progress (Sax, 1989).
Performance Assessment Test. In order to provide standards for performance assessment,
a participating expert teacher reviewed the written performance test. Moreover, an
27
academic research expert in science education, along with the researcher and the expert
teacher, agreed on the answers to each of the test questions. The test was also reviewed by
a reading specialist in the district for appropriateness and clarity of language. Finally, the
written test was piloted with a fifth-grade class to assess the level of difficulty in
interpreting and understanding the questions. Many changes and modifications were
made throughout this process in order to develop a written test that was appropriate,
reliable, and valid.
Concerning the nature and form of the test questions, several procedures exist for
assessing cognitive skills (Royer, Cisero, & Carlo, 1993). However, the previously
described techniques used to assess the use of structural knowledge and higher order
cognitive skills (Jonassen, Beissner, & Yacci, 1993; Dierkhoff , 1983), have been
successfully utilized with hypertext software experiments (Jonassen & Wang, 1993).
The method seems also to be best suited for this research study.
Data Collection
Almost all the research data associated with the previously described experimental
variables (cognitive style, knowledge structure, and written performance test) were
28
obtained during a three week period. However, the CTBS reading percentile scores data
were readily available from the student records, since a version of this test (Grade 4.5
Level) was administered during the previous school year (Spring, 1995). The CTBS test
is administered yearly by the district to all enrolled fourth-graders to determine whether a
student needs remedial help in a particular subject.
RESULTS
All data were collected, with cooperation of the schools and participating teachers.
Observations were performed in the computer classrooms throughout the experiment.
Some students were also randomly chosen for brief interviews. The development of the
three ecology hyperdocuments used in the study was a much longer process that spanned
many months. This process also included cooperation and input from participating
teachers and domain experts.
Data was also gathered on the student subjects' reading levels by using their CTBS
scores (CTB/MacMillan, 1989). The CTBS test is administered yearly by the district to
all enrolled fourth grade students. A version of this test (Grade 4.5 Level) was
administered to all fourth-graders during the previous school year. Therefore, the CTBS
reading scores (in terms of percentiles) were readily available from the school records for
those students
29
enrolled in the district the previous year. Student subjects classified as "special education"
or "learning disabled" were removed from all research analyses.
Research Hypotheses
The following are the four hypotheses identified and addressed by this research study:
1. There will be a significant relationship between the learner's cognitive style
(active or reflective) and his/her performance, when using a hypermedia learning system.
2. There will be a significant relationship between the learner's usage of the
knowledge structure schema reflected by the hypermedia learning system (hierarchical,
branching, or conventional), and his/her performance.
3. There will be a significant relationship between the learner's usage of the
knowledge structure schema reflected by the hypermedia learning system (hierarchical,
branching, or conventional), and his/her performance in terms of higher-order cognitive
skills.
4. There will be a significant interaction between the learner's cognitive style
and the usage of the knowledge structure schema reflected by the hypermedia learning
system, in relation to his/her performance.
30
Analysis of Covariance
The research data described above were examined by using the analysis of
covariance (ANCOVA) statistical technique. In all cases, homogeneities of variance and
regression were shown. An ANCOVA was performed for all three performance scores:
Total, Higher Cognitive Skills, and Lower Cognitive Skills, with Reading as the covariate
and a sample population of 115.
Table 1 presents the ANCOVA results for the Total Performance dependent
variable, with Knowledge Structure (three groups) and Cognitive Style (two groups) as
factors. A significance level of p=.05 was selected. As can be observed, the results
showed a significant difference among the Knowledge Structure groups (F = 3.31, p <
.05), but no significant difference between the Cognitive Style groups (F = .70, p < .05)
was exhibited. Furthermore, no interaction between Knowledge Structure and Cognitive
Style was observed (F = .17, p < .05).
[Insert Table 1 here]
In further exploring the statistically significant results associated with the
Knowledge Structure groups, a post-hoc test was performed using the Student-NewmanKeuls procedure (Huck, Cormier, & Bounds, 1974). Once again, a significance level of
p=.05 was selected. A significant difference between the Branching and Hierarchical
31
groups (F = 5.58, p < .05), and between the Branching and Conventional groups
(F = 6.68, p < .05) was found. No significant difference was found between the
Hierarchical and Conventional groups.
Similarly, Table 2 shows the ANCOVA results for the Higher Cognitive Skills
Performance dependent variable, with Knowledge Structure and Cognitive Style as
factors. A significant difference among the Knowledge Structure groups (F = 4.05,
p < .05) was found, but no significant difference between the Cognitive Style groups
(F = .27, p < .05) was exhibited. Furthermore, no interaction between Knowledge
Structure and Cognitive Style was observed.
[Insert Table 2 here]
A Student-Newman-Keuls procedure was used to further analyze the
Knowledge Structure groups at a significance level of p=.05. This test showed a
significant difference between the Branching and Conventional groups (F = 8.59, p <
.05). However, no significant difference was found between the Hierarchical and
Conventional groups and the Hierarchical and Branching groups.
Finally, Table 3 shows the ANCOVA results for the Lower Cognitive Skills
Performance dependent variable, with Knowledge Structure and Cognitive Style as
factors. The results showed no significant difference among the Knowledge Structure
groups and between the Cognitive Style groups. Furthermore, no interaction between
Knowledge Structure and Cognitive Style was observed.
32
[Insert Table 3 here]
FINDINGS
The objective of this research study has been to analyze the effects that the
structuring of the knowledge domain (as reflected by the use of three major schemas), and
the cognitive style of the learner had on performance when using a hypermedia learning
system. Six fifth-grade classrooms from three different elementary schools within the
same school district were chosen for the study. Each of the three hyperdocuments,
specifically designed for this experiment, was randomly assigned to two classrooms. The
domain of ecology was carefully chosen and used in this study to control for the students'
prior knowledge.
All subjects used their assigned hyperdocument system on two different
occasions, over a period of a week, in a computer classroom. A total of 115 students
were represented in the experimental sample; 44 students in the hierarchical group, 38
students in the branching group, and 33 students in the conventional group.
Along with a cognitive style instrument, the students were also administered an
immediate written posttest that assessed their learning performance in terms of higher and
33
lower-order cognitive skills. The performance results were analyzed to determine the
effects that the cognitive style and knowledge structure factors had upon the learner.
An analysis of covariance was used to control for the initial differences of reading
ability that might have existed among the three groups. The differences among the group
means were used to test the related hypotheses.
The following is a summary of the findings of the research study, regarding the
effects that the structuring of the knowledge domain and the cognitive style of the learner
have on performance when using a hypermedia learning system.
Hypothesis 1. The first hypothesis tested was that there would be a significant
relationship between the learner's cognitive style (active or reflective) and his/her
performance, when using a hypermedia learning system. No significant difference was
observed with any of the three performance mean scores for the two types of styles
considered in the study (active and reflective). Although the active styles group scored
slightly higher (M = 24.11) than the reflective styles group (M = 23.75) in both total
performance and higher cognitive skills performance, these differences were not
significant. On the other hand, the reflective styles group scored slightly higher
(M = 11.71) than the active styles group (M = 11.60) with the lower cognitive skills
performance. Once again, these differences were not significant.
34
Hypothesis 2. The second tested hypothesis was that there would be a significant
relationship between the learner's usage of the knowledge structure schema reflected by
the hypermedia learning system (hierarchical, branching, or conventional), and his/her
performance. Significant differences were found with the total performance scores among
the three knowledge structure schema groups (F = 3.31). Subsequent post-hoc tests
revealed that the branching group performed significantly better (M = 25.55) than the
hierarchical group (M = 23.36) and the conventional group (M = 22.88), with F = 5.58
and F = 6.86, respectively.
Hypothesis 3. The third test hypothesis was that there would be a significant
relationship between the learner's usage of the knowledge structure schema reflected by
the hypermedia learning system (hierarchical, branching, or conventional), and his/her
performance in terms of higher order cognitive skills. A significant difference was found
with the higher cognitive skills mean performance scores among the three knowledge
structure schema groups (F = 4.05). A subsequent post-hoc test revealed that the
branching group performed significantly better (M = 13.32) than the conventional group
(M = 11.21) but not significantly better than the hierarchical group (M = 12.23),
with F = 8.59 and F = 2.82, respectively. Furthermore, no significant relationship was
detected among the lower cognitive skills mean performance scores of the knowledge
structure groups.
35
Hypothesis 4. The fourth and final hypothesis tested was that there would be a
significant interaction between the learner's cognitive style and the usage of the
knowledge structure schema reflected by the hypermedia learning system, in relation to
his/her performance. No significant interaction was observed between the two
independent variables (cognitive style and knowledge structure) with any of the three
learner performance scores.
DISCUSSION
The use of hypermedia software as learning systems does not necessarily lead to
improved performance. As previously discussed, a high level of learner control, coupled
with the rich learning environment provided by the hypermedia system, can lead to
distraction and "hyperchaos". Under these circumstances, it is highly probable that
learning will not occur, since students can either miss relevant instructional points, and
perhaps form wrong interpretations of the information. Strategies for adding structure to
hyperdocuments, such as those employed by this experiment, have been recommended to
minimize this problem. The research findings seem to indicate that, the exploratory or
navigational freedom associated with the knowledge structure schemas chosen for this
36
experiement (hierarchical, branching, and conventional) did not distract the student
subjects from the learning task at hand.
Thus, the research findings of this study show that properly structured hypermedia
software can be effectively used as learning systems to improve performance and
achievement. Moreover, it appears that the hypermedia systems provided an effective
means for promoting and developing the learner's higher cognitive skills (e.g., analysis
and synthesis).
The knowledge structure schemas used in the study (hierarchical, branching, and
conventional), were chosen to provide a range of degrees of navigational or exploratory
freedom. It seems evident from the findings that when "too much" freedom (as in the case
of the conventional schema) or "not enough" freedom (as in the case of the hierarchical
schema) is provided to early adolescent students in the form of a computer learning
system, performance may suffer. This fact was evidenced by both the quantitative
data analyzed above, and the numerous personal observations and student interviews
conducted during the course of the study.
Although initially very enthusiastic, it was interesting to note that many student
subjects became either easily bored (as was the case with the hierarchical group) or
turned the learning system into a game (as was the case with the conventional group).
Since the hierarchical knowledge schema allowed the subjects to easily visit all of the
system screens in a relatively short period of time, many students in this group became
37
quickly bored after they had viewed and heard all the "cool" images, videos, and sounds.
Even after repeated instructions from the teachers to read all the textual information,
many appeared bored and restless.
However, with the conventional knowledge schema group, many students became
intrigued with locating and replaying their favorite screen(s) (image, sound, or video).
Since the conventional hyperdocument did not allow for this task to be easily
accomplished, this became a challenging game for many students. Once again, even after
repeated instructions from the teachers to read the textual information, many continued
along with this form of behavior. On the other hand, the members of this group seemed
much more engaged, and not nearly as bored as the hierarchical group participants.
Finally, the branching hyperdocument seemed to have offered its group members
the appropriate freedom to navigate the knowledge space and enough structure to allow
them to visit all the screens. The level of structure offered by this schema seemed to
have benefited the students by holding their interest and focusing their attention on the
task. Consequently, these observations, along with the data results, indicate that the
branching hyperdocument group performed significantly better than the others because
they were more engaged and focused (i.e., students did not become bored or distracted).
This finding also forms the basis for explaining the results associated with the
branching group performing significantly better than the conventional group in the higher
cognitive skills performance sub-test. It seems that an appropriate level of structure and
38
exploratory freedom may be required to promote higher cognitive skills. If the exploration
space becomes too large, the opportunity for getting distracted and disoriented increases.
This seems to be especially the case for the age group addressed by this study (early
adolescence), since their attention span has been deemed to be notoriously short. This
observation may also hold if the exploration space is too confining. Since the ability to
focus and concentrate on semantic relationships are required to promote higher cognitive
skills (Jonassen & Wang), a knowledge structure such as that offered by the branching
schema may be demanded.
In considering the effects of the student's cognitive styles on performance results,
no significant relationships could be established. More specifically, in examining the
information processing styles of Active Experimentation and Reflective Observation,
neither one could be significantly related to learner performance. Although not part
of the study, an analysis of the information perception styles dimension (Concrete
Experience and Abstract Conceptualization) yielded similarly non-significant results.
Furthermore, no significant interaction was observed between the cognitive styles
variable used in this study and the knowledge structure variable.
However, it is worthwhile to note that, when the research data were analyzed by
gender (male or female), "almost" significant relationships at the .05 level were observed
with male subjects for both the total (p < .07) and lower cognitive skills (p < .06)
39
performance scores. These inconclusive findings seem to add to the growing research
literature that have consistently shown mixed results in relating cognitive style and
learner performance (Paolucci & Wright, 1996).
This research study has clearly shown that hypermedia software, when
appropriately structured for the learner, can be effectively used as learning systems to
improve performance and achievement. Moreover, these learning systems provide an
effective means for promoting and developing the learner's higher cognitive skills.
Since no previous research had been conducted with adolescent students, this study
addressed this need by conducting an empirical study with early adolescent students.
In general, the research results imply that teachers and educators need to be
careful and pay close attention to how hypermedia software is applied in the curriculum.
The technology has tremendous flexibility in providing dynamic learning environments
for all types of students and needs. However, this flexibility seems to be a double-edged
sword. It seems that if the hypermedia learning environment is designed to be too
structured, many students will loose interest and the learning objectives may not be
achieved. On the other hand, if this environment is too unstructured, many students will
become confused and loose interest.
Thus, choosing the "right" structure seems to be the key in making effective use of
hypermedia learning systems in the learning process. This is probably even more
40
important for young students (e.g., adolescents), who perhaps require more structure with
their learning environments, and may need to be scaffolded throughout this early learning
stage.
The notion of "scaffolding" the learning process seems to be particularly
appropriate when using hypermedia. It may be an appropriate learning strategy to
decrease the level of structure of the hypermedia learning system as the student matures.
Culminating at the stage where the student himself or herself can actually develop their
own hypermedia document and construct knowledge (as many constructivists advocate).
The important point to be made here is that, this process will be highly dependent on the
individual's skills and knowledge base, thus any strategy should be individualized.
Teachers are encouraged to assess their students cognitive abilities (e.g., learning
styles) in order to design instructional strategies for optimal learning.
Currently, the trend in many classrooms seems to be in letting most students
create their own hypermedia documents. Thus, the above guidelines are rarely applied.
Although these hypermedia projects may be deemed to be "cool" and fun by the students
and teachers, very rarely is the question of whether knowledge of the domain has been
acquired or learning has taken place.
Similarly, with the explosive growth and popularization of the Internet's World
Wide Web, many teachers and educators (even at the elementary levels) are using these
41
highly hypermediated tools as learning environments. As this research study has shown, if
the knowledge domain is too unstructured (as is definitely the case with the World Wide
Web), learning becomes very difficult, especially for younger students. Therefore,
structuring the World Wide Web, or any other similarly hypermediated technology (e.g.,
CD-ROMs), to meet the cognitive skills and knowledge base requirements of the
learner, is critical.
In addition to the above observations, instructional systems designers and authors
need to be keenly aware of the proper application of multimedia information. In authoring
the HyperEcology software, it became very clear that by far the textual information was
the key medium in the knowledge transfer design process. Although the sounds, images,
and videos used in the hyperdocuments were chosen to enhance learning, most of this
information was contained within the context of textual information base.
During the experiment, it soon became evident that the major focus of attention
for the majority of the students were the sounds, images, and videos. Consequently, much
textual information may not have been fully processed by these students. It seems that,
although multimedia information may attract the attention of the learner (especially the
younger ones), and perhaps enhance the learning of new knowledge, textual information
still remains the key component in the transfer of knowledge with hypermedia. Thus,
designers need to make sure that if multimedia information is used (as it should be) in
authoring hypermedia, it is used to compel the learner to also read the text.
42
Finally, it is worthwhile to note that, quite a few students had difficulty
concentrating and focusing on the task at hand. These individuals were often observed to
be unable to sit still (of course, there were also many students that took the time to
methodically read the screens and tried to process the information, regardless of the
hyperdocument used). No technology will be effective in enhancing knowledge
acquisition if attention and focus is not given. Thus, whether hypermedia can be effective
for these students is questionable. On the contrary, unfortunately, it may even exacerbate
this condition by providing an overly stimulating learning environment.
References
Ambrose, D. (1991). The effects of hypermedia on learning. Educational Technology,
December, 51-55.
Becker, H. J. (1993). Teaching with and about computers in secondary schools.
Communications of the ACM, 36 (1), 69-73.
Bevilacqua, A. F. (1989). Hypertext: Behind the hype. American Libraries, 20, 158-162.
Billings, D. M., & Cobb, K. L. (1992). Effects of learning style, preferences, attitude, and
GPA on learner achievement using computer assisted interactive videodisc
instruction. Journal of Computer-Based Instruction, 19 (1), 12-16.
43
Bloom, B. S. (1956). Taxonomy of educational objectives: The classification of
educational goals, Handbook I: Cognitive domain. New York, NY: Longman &
Green.
Borg, W. R., & Gall, M. D. (1989). Educational research: An introduction, (5th ed.).
White Plains, NY: Longman.
Brooks, L. W., Simutis, Z. M., & O'Neil, H. F. (1985). The role of individual differences
in learning strategies. In R. F. Dillon, (Ed.), Individual differences and cognition,
v.2. U.S.A.: Academic Press.
Bush, V. (1945). As we may think. Atlantic Monthly, 176, 101-108.
CTB/MacMillan (1989). Comprehensive Tests of Basic Skills, (4th ed.). Monterey, CA:
McGraw-Hill.
Cates, W. M. (1992). Fifteen principles for designing more effective instructional
hypermedia/multimedia products. Educational Technology, December, 5-11.
Clabaugh, G. K., & Rozycki, E. G. (1990). Understanding schools. New York:
Harper & Row.
Conklin, J. (1987). Hypertext: An introduction and survey. Computer, 20 (9), 17-41.
Cordell, B.J. (1991). A study of learning styles and computer-assisted instruction.
Computers in Education, 16 (2), 175-183.
Dede, C. J. (1992). The future of multimedia: Bridging to virtual worlds. Educational
Technology, 32 (5), 54-60.
44
Dern, D. P. (1993). Applying the internet. Byte, 17 (2), 111-114.
Dierkhoff, G. M. (1983). Testing through relationship judgments. Journal of Educational
Psychology, 75, 227-233.
Dunn, R., Dunn, K., & Price, G. E. (1984). Learning style inventory. Lawrence, KS: Price
Systems.
Englebart, D. (1963). A conceptual framework for augmentation of man's intellect. In P.
Howerton, & D. Weeks, (Eds.), Vistas in information handling, Vol. 1.
Washington, DC: Spartan Books.
Gagne, R. M. (1987). The conditions of learning, (4th ed.). New York: Holt, Rinehart,
and Winston.
Hannafin, M. J. (1992). Emerging technologies, ISD, and learning environments: Critical
perspectives. Educational Technology Research & Development, 40 (1), 49-63.
Hawkins, J. (1993). Technology and the organization of schooling. Communications of
the ACM, 36 (5), 30-34.
Huck, S.W., Cormier, W.H., & Bounds, W.G., Jr., (1974). Reading statistics and
research. New York: Harper Collins.
Hunter, B. (1993). Internetworking: Coordinating technology for system reform.
Communications of the ACM, 36 (1), 42-46.
45
Jonassen, D. (1988). Designing structured hypertext and structuring access to hypertext.
Educational Technology, November, 13-16.
Jonassen, D. (1991). Evaluating constructivist learning. Educational Technology,
September, 28-33.
Jonassen, D. (1993). Conceptual frontiers in hypermedia environments for learning.
Journal of Educational Multimedia and Hypermedia, 2 (4), 331-335.
Jonassen, D., Beissner, K, & Yacci, M. A. (1993). Structural knowledge: Techniques for
assessing, conveying and acquiring structural knowledge. Hillsdale, NJ:
Lawrence Earlbaum Associates.
Jonassen D., & Wang, S. (1993). Acquiring structural knowledge from semantically
structured hypertext. Journal of Computer-Based Instruction, 20 (1), 1-8.
Kearsley, G. (1988). Authoring considerations for hypertext. Educational Technology,
November, 21-24.
Keefe, J. W. (1988). Profiling and utilizing learning style. Reston, VA: National
Association of Secondary School Principals.
Kolb, D. A. (1976). Learning style inventory technical manual. Boston: McBer & Co.
Kolb, D. A. (1984). Experiential learning: Experience as the source of learning
and development. Englewood Cliffs, NJ: Prentice-Hall.
Kumar, D. (1994). Hypermedia: A tool for alternative assessment? Educational
and Training Technology International, 31 (1), 59-67.
46
Larsen, R. (1992). Relationship of learning styles to the effectiveness and acceptance of
interactive video instruction. Journal of Computer-Based Instruction, 19 (1),
17-21.
Locatis, C., Letourneau, G., & Banvard, R. (1992). Hypermedia and instruction.
Educational Technology Research and Development, 37 (4), 65-77.
Marchionini, G. (1988). Hypermedia and learning: Freedom and chaos. Educational
Technology, November, 8-12.
McCarthy, B. (1981). The 4Mat system: Teaching to learning styles with right/left
mode techniques. Oak Brook, IL: EXCEL, Inc.
Merrill, M. D. (1983). Component display theory. In C. M. Reigeluth (Ed.), Instructional
theories and models. Hillsdale, NJ: Erlbaum.
Misanchuk, E. R., & Schwier, R. (1991). Interactive media audit trails: Approaches and
issues. ERIC Abstracts, (ED 334996).
Morariu, J. (1988). Hypermedia in instruction and training: The power and the promise.
Educational Technology, November, 17-20.
Nelson, T. H. (1974). Dream machines. South Bend, IN: The Distributors.
Nelson, W. (1992). Adaptive hypermedia instructional systems: Possibilities for learner
modeling. ERIC Abstracts, (ED 347195).
47
Oren, T. (1990). Cognitive load in hypermedia: Designing for the exploratory learner. In
S. Ambron, & K. Hooper, (Eds.), Learning with interactive multimedia.
Redmond, WA: Microsoft Press.
Paolucci, R., & Wright, R.J. (1996). Hypermedia learning environments: A research
review of the effects of cognitive style on performance. Proceedings of the World
Conference on Educational Multimedia and Hypermedia, USA, 858.
Park, O. (1991). Hypermedia: Functional features and research issues. Educational
Technology, August, 1991, 24-31.
Pearlman, R. (1993). Designing the new American schools. Communications of
the ACM, 36 (1), 46-49.
Rasmussen, K., & Davidson, G.V. (1996). Dimensions of learning styles and their
influence on performance in hypermedia lessons. Proceedings of the World
Conference on Educational Multimedia and Hypermedia, USA, 800.
Royer, J. M., Cisero, C. A., & Carlo, M. S. (1993). Techniques and procedures for
assessing cognitive skills. Review of Educational Research, 63 (2), 201-243.
Sax, G. (1989). Principles of educational and psychological measurement and
evaluation, (3rd ed.). Belmont, CA: Wadsworth.
Schank, R. C. (1993). Learning via multimedia computers. Communications of the ACM,
36 (1), 54-56.
48
Sheingold, K. (1991). Restructuring for learning with technology: The potential for
synergy. Phi Delta Kappan, 73 (1), 17-27.
Shin, C., Schallert, D., & Savenye, W. (1994). Effects of learner control, advisement, and
prior knowledge on young students' learning in a hypertext environment.
Educational Technology Research and Development, 42 (1), 33-47.
Smith, D. M., & Kolb, D. A. (1986). User's guide for the learning style inventory.
Boston: McBer & Co.
Snider, R. C. (1992). The machine in the classroom. Phi Delta Kappan, 74 (4), 316-323.
Soloway, E. (1993). Reading and writing in the 21st century. Communications of the
ACM, 36 (1), 23-27.
Soloway, E. (1995). Beware, techies bearing gifts. Communications of the ACM, 38 (1),
17-24.
Stanton, N. A. (1994). Explorations into hypertext: Spatial metaphorconsidered harmful.
Educational and Training Technology International, 31 (4), 277-295.
Tolhurst, D. (1992). A checklist for evaluating content-based hypertext computer
software. Educational Technology, March, 17-20.
Tsai, C. (1989). Hypertext: technology, applications, and research issues. Journal of
Educational Technology Systems, 17 (1), 3-15.
Vandergrift, K. E. (1988). Hypermedia: Breaking the tyranny of the text. School Library
Journal, 35, 30-35.
49
Wagner, R. (1995). HyperStudio Reference. El Cajon, CA: Roger Wagner Publishing.
Whyte, M. M., Karolick, D. M., Nielsen, M. C., Elder, G. D., & Hawley, W. T. (1995).
Cognitive styles and feedback in computer-assisted instruction. Journal of
Educational Computing Research, 12 (2), 195-203.
Yung-Bin, B. L. (1992). Effects of learning style in a hypermedia instructional system.
ERIC Abstracts, (ED 348008)
Table 1
Analysis of Covariance of Total Performance Scores
Source
Knowledge Structure (A)
Cognitive Style (B)
Reading (Covar)
AxB
Within Error
Total
*
**
SS
116.34
12.31
339.62
6.07
1667.87
2142.21
DF
2
1
1
2
95
101
MS
58.17
12.31
339.62
3.03
17.56
F
3.31*
.70
19.34**
.17
p<.05
p<.001
Table 2
Analysis of Covariance of Higher-Order Cognitive Skills Performance Scores
Source
Knowledge Structure (A)
Cognitive Style (B)
Reading (Covar)
AxB
Within Error
Total
*
**
SS
74.94
2.51
90.55
6.14
879.44
1053.58
DF
2
1
1
2
95
101
MS
37.47
2.51
90.55
3.07
9.26
F
4.05*
.27
9.78**
.33
p<.05
p<.001
Table 3
Analysis of Covariance of Lower-Order Cognitive Skills Performance Scores
Source
Knowledge Structure (A)
Cognitive Style (B)
Reading (Covar)
AxB
Within Error
Total
**
p<.001
SS
18.36
3.71
79.44
.20
440.60
542.31
DF
2
1
1
2
95
101
MS
9.18
3.71
79.44
.10
4.64
F
1.98
.80
17.13**
.02